4.

1 PRELIMINARY THEORY—LINEAR EQUATIONS ● 117

4.1 PRELIMINARY THEORY—LINEAR EQUATIONS

REVIEW MATERIAL
● Reread the Remarks at the end of Section 1.1

● Section 2.3 (especially page 57)

INTRODUCTION In Chapter 2 we saw that we could solve a few first-order differential equa-
tions by recognizing them as separable, linear, exact, homogeneous, or perhaps Bernoulli equations.
Even though the solutions of these equations were in the form of a one-parameter family, this
family, with one exception, did not represent the general solution of the differential equation. Only
in the case of linear first-order differential equations were we able to obtain general solutions, by
paying attention to certain continuity conditions imposed on the coefficients. Recall that a general
solution is a family of solutions defined on some interval I that contains all solutions of the DE that
are defined on I. Because our primary goal in this chapter is to find general solutions of linear higher-
order DEs, we first need to examine some of the theory of linear equations.

4.1.1 INITIAL-VALUE AND BOUNDARY-VALUE

PROBLEMS
Initial-Value Problem In Section 1.2 we defined an initial-value problem for
a general nth-order differential equation. For a linear differential equation an
nth-order initial-value problem is

d ny d n1y dy
Solve: an(x)  an1(x)      a1(x)  a0(x)y  g(x)
dx n dx n1 dx (1)
Subject to: y(x0)  y0, y(x0)  y1 , . . . , y(n1)(x0)  yn1.

Recall that for a problem such as this one we seek a function defined on some interval
I, containing x 0 , that satisfies the differential equation and the n initial conditions
specified at x 0: y(x 0 )  y 0 , y(x 0 )  y1, . . . , y (n1)(x 0)  y n1. We have already seen
that in the case of a second-order initial-value problem a solution curve must pass
through the point (x 0 , y 0) and have slope y1 at this point.

Existence and Uniqueness In Section 1.2 we stated a theorem that gave

conditions under which the existence and uniqueness of a solution of a first-order
initial-value problem were guaranteed. The theorem that follows gives sufficient
conditions for the existence of a unique solution of the problem in (1).

THEOREM 4.1.1 Existence of a Unique Solution

Let a n(x), a n1(x), . . . , a1(x), a 0(x) and g(x) be continuous on an interval I and
let a n(x)  0 for every x in this interval. If x  x 0 is any point in this interval,
then a solution y(x) of the initial-value problem (1) exists on the interval and is
unique.

EXAMPLE 1 Unique Solution of an IVP

The initial-value problem

3y   5y  y  7y  0, y(1)  0, y(1)  0, y (1)  0

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
118 ● CHAPTER 4 HIGHER-ORDER DIFFERENTIAL EQUATIONS

possesses the trivial solution y  0. Because the third-order equation is linear with
constant coefficients, it follows that all the conditions of Theorem 4.1.1 are fulfilled.
Hence y  0 is the only solution on any interval containing x  1.

EXAMPLE 2 Unique Solution of an IVP

You should verify that the function y  3e 2x  e2x  3x is a solution of the initial-
value problem
y  4y  12x, y(0)  4, y(0)  1.
Now the differential equation is linear, the coefficients as well as g(x)  12x are
continuous, and a2(x)  1  0 on any interval I containing x  0. We conclude from
Theorem 4.1.1 that the given function is the unique solution on I.

The requirements in Theorem 4.1.1 that a i (x), i  0, 1, 2, . . . , n be continuous

and a n (x)  0 for every x in I are both important. Specifically, if a n (x)  0 for some x
in the interval, then the solution of a linear initial-value problem may not be unique
or even exist. For example, you should verify that the function y  cx 2  x  3 is a
solution of the initial-value problem
x2 y  2xy  2y  6, y(0)  3, y(0)  1

on the interval ( , ) for any choice of the parameter c. In other words, there is no
unique solution of the problem. Although most of the conditions of Theorem 4.1.1
are satisfied, the obvious difficulties are that a2(x)  x 2 is zero at x  0 and that the
initial conditions are also imposed at x  0.

Boundary-Value Problem Another type of problem consists of solving a lin-

solutions of the DE
y ear differential equation of order two or greater in which the dependent variable y or
its derivatives are specified at different points. A problem such as

d 2y dy
Solve: a2(x) 2
 a1(x)  a0(x)y  g(x)
dx dx
(b, y1) Subject to: y(a)  y0 , y(b)  y1
(a, y0)
is called a boundary-value problem (BVP). The prescribed values y(a)  y0 and
x y(b)  y1 are called boundary conditions. A solution of the foregoing problem is a
I function satisfying the differential equation on some interval I, containing a and b,
FIGURE 4.1.1 Solution curves of a whose graph passes through the two points (a, y0) and (b, y1). See Figure 4.1.1.
BVP that pass through two points For a second-order differential equation other pairs of boundary conditions
could be

y(a)  y0 , y(b)  y1
y(a)  y0 , y(b)  y1
y(a)  y0 , y(b)  y1,

where y0 and y1 denote arbitrary constants. These three pairs of conditions are just
special cases of the general boundary conditions

1 y(a)  1 y(a)  1

2 y(b)  2 y(b)  2.

The next example shows that even when the conditions of Theorem 4.1.1 are
fulfilled, a boundary-value problem may have several solutions (as suggested in
Figure 4.1.1), a unique solution, or no solution at all.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 4.1 PRELIMINARY THEORY—LINEAR EQUATIONS ● 119

EXAMPLE 3 A BVP Can Have Many, One, or No Solutions

In Example 7 of Section 1.1 we saw that the two-parameter family of solutions of the
differential equation x  16x  0 is
x  c1 cos 4t  c2 sin 4t. (2)

x
(a) Suppose we now wish to determine the solution of the equation that further
c2 = 1 satisfies the boundary conditions x(0)  0, x(p兾2)  0. Observe that the first
1
c2 = 2
1
1 condition 0  c1 cos 0  c2 sin 0 implies that c1  0, so x  c2 sin 4t. But when t 
c2 =
c2 = 0
4 p兾2, 0  c2 sin 2p is satisfied for any choice of c2, since sin 2p  0. Hence the
boundary-value problem
t

(0, 0)
c2 = −
1
(π /2, 0) x  16x  0, x(0)  0, x 冢2 冣  0 (3)
1 2
has infinitely many solutions. Figure 4.1.2 shows the graphs of some of the
FIGURE 4.1.2 Solution curves for members of the one-parameter family x  c2 sin 4t that pass through the two points
BVP in part (a) of Example 3 (0, 0) and (p兾2, 0).
(b) If the boundary-value problem in (3) is changed to

x  16x  0, x(0)  0, x 冢8 冣  0, (4)

then x(0)  0 still requires c1  0 in the solution (2). But applying x(p兾8)  0 to
x  c 2 sin 4t demands that 0  c 2 sin(p兾2)  c2  1. Hence x  0 is a solution of this
new boundary-value problem. Indeed, it can be proved that x  0 is the only solution
of (4).
(c) Finally, if we change the problem to

x  16x  0, x(0)  0, x 冢2 冣  1, (5)

we find again from x(0)  0 that c1  0, but applying x(p兾2)  1 to x  c2 sin 4t leads
to the contradiction 1  c2 sin 2p  c2  0  0. Hence the boundary-value problem (5)
has no solution.

4.1.2 HOMOGENEOUS EQUATIONS

A linear nth-order differential equation of the form

dny d n1y dy
an(x)  an1(x)      a1(x)  a0(x)y  0 (6)
dx n dx n1 dx
is said to be homogeneous, whereas an equation

dny d n1y dy
an(x) n
 an1(x) n1      a1(x)  a0(x)y  g(x), (7)
dx dx dx
with g(x) not identically zero, is said to be nonhomogeneous. For example,
2y  3y  5y  0 is a homogeneous linear second-order differential equation,
whereas x 3y  6y  10y  e x is a nonhomogeneous linear third-order differen-
tial equation. The word homogeneous in this context does not refer to coefficients
that are homogeneous functions, as in Section 2.5.
We shall see that to solve a nonhomogeneous linear equation (7), we must first
be able to solve the associated homogeneous equation (6).
To avoid needless repetition throughout the remainder of this text, we
shall, as a matter of course, make the following important assumptions when

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
120 ● CHAPTER 4 HIGHER-ORDER DIFFERENTIAL EQUATIONS

stating definitions and theorems about linear equations (1). On some common
interval I,
Please remember
these two
䉴 • the coefficient functions a i (x), i  0, 1, 2, . . . , n and g(x) are continuous;
assumptions.
• a n(x)  0 for every x in the interval.

Differential Operators In calculus differentiation is often denoted by the cap-

ital letter D —that is, dy兾dx  Dy. The symbol D is called a differential operator
because it transforms a differentiable function into another function. For example,
D(cos 4x)  4 sin 4x and D(5x 3  6x 2 )  15x 2  12x. Higher-order derivatives
can be expressed in terms of D in a natural manner:
d dy
dx dx 冢 冣
d 2y
 2  D(Dy)  D2y
dx
and, in general,
dny
dxn
 Dn y,

where y represents a sufficiently differentiable function. Polynomial expressions

involving D, such as D  3, D 2  3D  4, and 5x 3D 3  6x 2D 2  4xD  9, are
also differential operators. In general, we define an nth-order differential opera-
tor or polynomial operator to be
L  an(x)D n  an1(x)D n1      a1(x)D  a 0 (x). (8)
As a consequence of two basic properties of differentiation, D(cf (x))  cDf (x), c is a
constant, and D{f (x)  g(x)}  Df (x)  Dg(x), the differential operator L possesses
a linearity property; that is, L operating on a linear combination of two differentiable
functions is the same as the linear combination of L operating on the individual func-
tions. In symbols this means that
L{a f (x)  bg(x)}  aL( f (x))  bL(g(x)), (9)
where a and b are constants. Because of (9) we say that the nth-order differential
operator L is a linear operator.

Differential Equations Any linear differential equation can be expressed in

terms of the D notation. For example, the differential equation y  5y  6y  5x  3
can be written as D2 y  5Dy  6y  5x  3 or (D 2  5D  6)y  5x  3. Using (8),
we can write the linear nth-order differential equations (6) and (7) compactly as
L(y)  0 and L(y)  g(x),
respectively.

Superposition Principle In the next theorem we see that the sum, or super-
position, of two or more solutions of a homogeneous linear differential equation is
also a solution.

THEOREM 4.1.2 Superposition Principle —Homogeneous Equations

Let y1, y2, . . . , yk be solutions of the homogeneous nth-order differential equation

(6) on an interval I. Then the linear combination
y  c1 y1(x)  c2 y2(x)      ck yk(x),
where the ci , i  1, 2, . . . , k are arbitrary constants, is also a solution on the
interval.

PROOF We prove the case k  2. Let L be the differential operator defined in

(8), and let y1(x) and y 2(x) be solutions of the homogeneous equation L( y)  0. If
we define y  c1 y1(x)  c 2 y 2(x), then by linearity of L we have

L( y)  L{c1 y1(x)  c2 y2(x)}  c1 L(y1)  c2 L(y2)  c1  0  c2  0  0.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 4.1 PRELIMINARY THEORY—LINEAR EQUATIONS ● 121

COROLLARIES TO THEOREM 4.1.2

(A) A constant multiple y  c1 y1(x) of a solution y1(x) of a homogeneous
linear differential equation is also a solution.
(B) A homogeneous linear differential equation always possesses the trivial
solution y  0.

EXAMPLE 4 Superposition—Homogeneous DE

The functions y1  x 2 and y2  x 2 ln x are both solutions of the homogeneous linear

equation x 3y  2xy  4y  0 on the interval (0, ). By the superposition principle
the linear combination
y  c1x2  c2 x2 ln x
is also a solution of the equation on the interval.

The function y  e7x is a solution of y  9y  14y  0. Because the differen-

tial equation is linear and homogeneous, the constant multiple y  ce7x is also a
solution. For various values of c we see that y  9e7x, y  0, y   15e7x, . . . are all
solutions of the equation.

Linear Dependence and Linear Independence The next two concepts

are basic to the study of linear differential equations.

DEFINITION 4.1.1 Linear Dependence/Independence

A set of functions f1(x), f2(x), . . . , fn(x) is said to be linearly dependent on an

interval I if there exist constants c1, c2, . . . , cn, not all zero, such that
c1 f1(x)  c2 f2(x)      cn fn(x)  0
for every x in the interval. If the set of functions is not linearly dependent on
the interval, it is said to be linearly independent.
y

f1 = x In other words, a set of functions is linearly independent on an interval I if the only

constants for which
x c1 f1(x)  c2 f2(x)      cn fn(x)  0
for every x in the interval are c1  c2      cn  0.
It is easy to understand these definitions for a set consisting of two functions
f1(x) and f 2(x). If the set of functions is linearly dependent on an interval, then
(a) there exist constants c1 and c2 that are not both zero such that for every x in the
interval, c1 f1 (x)  c2 f 2 (x)  0. Therefore if we assume that c1  0, it follows that
y f1 (x)  (c2 兾c1) f 2 (x); that is, if a set of two functions is linearly dependent, then one
function is simply a constant multiple of the other . Conversely, if f1(x)  c2 f 2(x)
f2 = |x| for some constant c2, then (1)  f1(x)  c2 f 2(x)  0 for every x in the interval.
Hence the set of functions is linearly dependent because at least one of the constants
x (namely, c1  1) is not zero. We conclude that a set of two functions f1(x) and f2(x)
is linearly independent when neither function is a constant multiple of the other on
the interval. For example, the set of functions f1(x)  sin 2x, f2(x)  sin x cos x is
linearly dependent on ( , ) because f1(x) is a constant multiple of f2(x). Recall
from the double-angle formula for the sine that sin 2x  2 sin x cos x. On the other
(b) hand, the set of functions f1(x)  x, f2(x)  兩 x兩 is linearly independent on ( , ).
FIGURE 4.1.3 Set consisting of f1 and Inspection of Figure 4.1.3 should convince you that neither function is a constant
f2 is linearly independent on ( , ) multiple of the other on the interval.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
122 ● CHAPTER 4 HIGHER-ORDER DIFFERENTIAL EQUATIONS

It follows from the preceding discussion that the quotient f2(x)兾f1(x) is not a con-
stant on an interval on which the set f1(x), f2(x) is linearly independent. This little fact
will be used in the next section.

EXAMPLE 5 Linearly Dependent Set of Functions

The set of functions f1(x)  cos 2 x, f2(x)  sin 2 x, f3(x)  sec 2 x, f4(x)  tan 2 x is
linearly dependent on the interval (p兾2, p兾2) because
c1 cos2x  c2 sin2x  c3 sec2x  c4 tan2x  0
when c1  c2  1, c3  1, c4  1. We used here cos2x  sin2x  1 and
1  tan2x  sec2x.

A set of functions f1(x), f2(x), . . . , fn(x) is linearly dependent on an interval if

at least one function can be expressed as a linear combination of the remaining
functions.

EXAMPLE 6 Linearly Dependent Set of Functions

The set of functions f1(x)  1x  5, f2(x)  1x  5x, f3(x)  x  1, f4(x)  x 2 is

linearly dependent on the interval (0, ) because f2 can be written as a linear combi-
nation of f1, f3, and f4. Observe that
f2(x)  1  f1(x)  5  f3(x)  0  f4(x)
for every x in the interval (0, ).

Solutions of Differential Equations We are primarily interested in linearly

independent functions or, more to the point, linearly independent solutions of a lin-
ear differential equation. Although we could always appeal directly to Definition 4.1.1,
it turns out that the question of whether the set of n solutions y1, y2 , . . . , yn of a
homogeneous linear nth-order differential equation (6) is linearly independent can be
settled somewhat mechanically by using a determinant.

DEFINITION 4.1.2 Wronskian

Suppose each of the functions f1(x), f2(x), . . . , fn(x) possesses at least n  1

derivatives. The determinant

兩 兩
f1 f2  fn
f 1 f 2  f n
W( f1, f2, . . . , fn )     ,
  
  
f1(n1) f2(n1)    fn(n1)
where the primes denote derivatives, is called the Wronskian of the
functions.

THEOREM 4.1.3 Criterion for Linearly Independent Solutions

Let y1, y 2 , . . . , yn be n solutions of the homogeneous linear nth-order

differential equation (6) on an interval I. Then the set of solutions is linearly
independent on I if and only if W(y1, y 2 , . . . , yn )  0 for every x in the
interval.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 4.1 PRELIMINARY THEORY—LINEAR EQUATIONS ● 123

It follows from Theorem 4.1.3 that when y1, y 2 , . . . , yn are n solutions of (6) on
an interval I, the Wronskian W( y1, y 2 , . . . , yn ) is either identically zero or never zero
on the interval.
A set of n linearly independent solutions of a homogeneous linear nth-order
differential equation is given a special name.

DEFINITION 4.1.3 Fundamental Set of Solutions

Any set y1, y 2 , . . . , y n of n linearly independent solutions of the homogeneous

linear nth-order differential equation (6) on an interval I is said to be a funda-
mental set of solutions on the interval.

The basic question of whether a fundamental set of solutions exists for a linear
equation is answered in the next theorem.

THEOREM 4.1.4 Existence of a Fundamental Set

There exists a fundamental set of solutions for the homogeneous linear nth-order
differential equation (6) on an interval I.

Analogous to the fact that any vector in three dimensions can be expressed as a
linear combination of the linearly independent vectors i, j, k, any solution of an nth-
order homogeneous linear differential equation on an interval I can be expressed as a
linear combination of n linearly independent solutions on I. In other words, n linearly
independent solutions y1, y 2 , . . . , yn are the basic building blocks for the general
solution of the equation.

THEOREM 4.1.5 General Solution — Homogeneous Equations

Let y1, y 2, . . . , yn be a fundamental set of solutions of the homogeneous linear nth-

order differential equation (6) on an interval I. Then the general solution of the
equation on the interval is
y  c1 y1(x)  c2 y2(x)      cn yn(x),
where ci , i  1, 2, . . . , n are arbitrary constants.

Theorem 4.1.5 states that if Y(x) is any solution of (6) on the interval, then con-
stants C1, C2, . . . , Cn can always be found so that
Y(x)  C1 y1(x)  C2 y2(x)      Cn yn(x).
We will prove the case when n  2.

PROOF Let Y be a solution and let y1 and y2 be linearly independent solutions of

a 2 y  a1 y  a 0 y  0 on an interval I. Suppose that x  t is a point in I for which
W(y1(t), y2(t))  0. Suppose also that Y(t)  k1 and Y(t)  k2. If we now examine
the equations
C1 y1(t)  C2 y2(t)  k1
C1 y1(t)  C2 y2(t)  k2,
it follows that we can determine C1 and C2 uniquely, provided that the determinant of
the coefficients satisfies

兩yy(t)(t)
1

1
y2(t)
y2(t) 兩
 0.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
124 ● CHAPTER 4 HIGHER-ORDER DIFFERENTIAL EQUATIONS

But this determinant is simply the Wronskian evaluated at x  t, and by assumption,

W  0. If we define G(x)  C1 y1(x)  C2 y2(x), we observe that G(x) satisfies the
differential equation since it is a superposition of two known solutions; G(x) satisfies
the initial conditions
G(t)  C1 y1(t)  C2 y2(t)  k1 and G(t)  C1 y1 (t)  C2 y2(t)  k2;
and Y(x) satisfies the same linear equation and the same initial conditions.
Because the solution of this linear initial-value problem is unique (Theorem 4.1.1),
we have Y(x)  G(x) or Y(x)  C1 y1(x)  C2 y2(x).

EXAMPLE 7 General Solution of a Homogeneous DE

The functions y1  e 3x and y2  e3x are both solutions of the homogeneous linear
equation y  9y  0 on the interval ( , ). By inspection the solutions are lin-
early independent on the x-axis. This fact can be corroborated by observing that the
Wronskian

兩3ee e3x
兩
3x
W(e3x, e3x )   6  0
3x
3e3x
for every x. We conclude that y1 and y2 form a fundamental set of solutions, and
consequently, y  c1e 3x  c2e3x is the general solution of the equation on the
interval.

EXAMPLE 8 A Solution Obtained from a General Solution

The function y  4sinh 3x  5e 3x is a solution of the differential equation in

Example 7. (Verify this.) In view of Theorem 4.1.5 we must be able to obtain this
solution from the general solution y  c1e 3x  c2e3x. Observe that if we choose
c1  2 and c2  7, then y  2e 3x  7e3x can be rewritten as

冢e  e3x
冣
3x
y  2e3x  2e3x  5e3x  4  5e3x.
2
The last expression is recognized as y  4 sinh 3x  5e3x.

EXAMPLE 9 General Solution of a Homogeneous DE

The functions y1  e x, y2  e 2x, and y3  e 3x satisfy the third-order equation

y  6y  11y  6y  0. Since
ex e2x e3x
W(e , e , e )  p ex 2e2x 3e3x p  2e6x  0
x 2x 3x

ex 4e2x 9e3x
for every real value of x, the functions y1, y2, and y3 form a fundamental set of solu-
tions on ( , ). We conclude that y  c1e x  c2e2x  c3e3x is the general solution
of the differential equation on the interval.

4.1.3 NONHOMOGENEOUS EQUATIONS

Any function yp, free of arbitrary parameters, that satisfies (7) is said to be a particular
solution or particular integral of the equation. For example, it is a straightforward
task to show that the constant function yp  3 is a particular solution of the
nonhomogeneous equation y  9y  27.

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
 4.1 PRELIMINARY THEORY—LINEAR EQUATIONS ● 125

Now if y1, y2 , . . . , yk are solutions of (6) on an interval I and yp is any particular

solution of (7) on I, then the linear combination

y  c1 y1 (x)  c2 y2(x)      ck yk(x)  yp (10)

is also a solution of the nonhomogeneous equation (7). If you think about it, this makes
sense, because the linear combination c1 y1(x)  c2 y2(x)      ck yk (x) is trans-
formed into 0 by the operator L  an D n  a n1D n1      a1 D  a 0 , whereas yp
is transformed into g(x). If we use k  n linearly independent solutions of the nth-order
equation (6), then the expression in (10) becomes the general solution of (7).

THEOREM 4.1.6 General Solution —Nonhomogeneous Equations

Let yp be any particular solution of the nonhomogeneous linear nth-order differen-

tial equation (7) on an interval I, and let y1, y2, . . . , yn be a fundamental set of so-
lutions of the associated homogeneous differential equation (6) on I. Then the
general solution of the equation on the interval is

y  c1 y1(x)  c2 y2(x)      cn yn(x)  yp ,

where the ci , i  1, 2, . . . , n are arbitrary constants.

PROOF Let L be the differential operator defined in (8) and let Y(x) and yp(x)
be particular solutions of the nonhomogeneous equation L(y)  g(x). If we define
u(x)  Y(x)  yp(x), then by linearity of L we have

L(u)  L{Y(x)  yp(x)}  L(Y(x))  L(yp(x))  g(x)  g(x)  0.

This shows that u(x) is a solution of the homogeneous equation L(y)  0. Hence by
Theorem 4.1.5, u(x)  c1 y1(x)  c2 y2(x)      cn yn(x), and so

Y(x)  yp(x)  c1 y1(x)  c2 y2(x)      cn yn(x)

or Y(x)  c1 y1(x)  c2 y2(x)      cn yn(x)  yp(x).

Complementary Function We see in Theorem 4.1.6 that the general solu-

tion of a nonhomogeneous linear equation consists of the sum of two functions:

y  c1 y1(x)  c2 y2(x)      cn yn(x)  yp(x)  yc(x)  yp(x).

The linear combination yc(x)  c1 y1(x)  c2 y2(x)      cn yn(x), which is the

general solution of (6), is called the complementary function for equation (7). In
other words, to solve a nonhomogeneous linear differential equation, we first solve
the associated homogeneous equation and then find any particular solution of the
nonhomogeneous equation. The general solution of the nonhomogeneous equation
is then

y  complementary function  any particular solution

 yc  yp.

EXAMPLE 10 General Solution of a Nonhomogeneous DE

By substitution the function yp  11 1

12  2 x is readily shown to be a particular solu-
tion of the nonhomogeneous equation

y  6y  11y  6y  3x. (11)

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.
126 ● CHAPTER 4 HIGHER-ORDER DIFFERENTIAL EQUATIONS

To write the general solution of (11), we must also be able to solve the associated
homogeneous equation
y  6y  11y  6y  0.
But in Example 9 we saw that the general solution of this latter equation on the in-
terval ( , ) was yc  c1e x  c2e2x  c3e3x. Hence the general solution of (11)
on the interval is
11 1
y  yc  yp  c1ex  c2e2x  c3e3x   x.
12 2

Another Superposition Principle The last theorem of this discussion will

be useful in Section 4.4 when we consider a method for finding particular solutions
of nonhomogeneous equations.

THEOREM 4.1.7 Superposition Principle — Nonhomogeneous Equations

Let yp1, yp2 , . . . , ypk be k particular solutions of the nonhomogeneous linear nth-
order differential equation (7) on an interval I corresponding, in turn, to k dis-
tinct functions g1, g2, . . . , gk. That is, suppose ypi denotes a particular solution
of the corresponding differential equation

an(x)y(n)  an1(x)y(n1)      a1(x)y  a0(x)y  gi (x), (12)

where i  1, 2, . . . , k. Then

yp  yp1(x)  yp2(x)      ypk(x) (13)

is a particular solution of

an(x)y(n)  an1(x)y(n1)      a1(x)y  a0(x)y

 g1(x)  g2(x)      gk(x). (14)

PROOF We prove the case k  2. Let L be the differential operator defined in (8)
and let yp1(x) and yp2(x) be particular solutions of the nonhomogeneous equations
L( y)  g1(x) and L( y)  g2(x), respectively. If we define yp  yp1(x)  yp2(x), we
want to show that yp is a particular solution of L( y)  g1(x)  g2(x). The result
follows again by the linearity of the operator L:
L(yp)  L{yp1(x)  yp2(x)}  L( yp1(x))  L( yp2(x))  g1(x)  g2(x).

EXAMPLE 11 Superposition —Nonhomogeneous DE

You should verify that

yp1  4x2 is a particular solution of y  3y  4y  16x2  24x  8,
yp2  e2x is a particular solution of y  3y  4y  2e2x,
yp3  xex is a particular solution of y  3y  4y  2xex  ex.
It follows from (13) of Theorem 4.1.7 that the superposition of yp1, yp2, and yp3,
y  yp1  yp2  yp3  4x2  e2x  xex,
is a solution of

y  3y  4y  16x2  24x  8  2e2x  2xex  ex.

g1(x) g2(x) g3(x)

Copyright 2012 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part. Due to electronic rights, some third party content may be suppressed from the eBook and/or eChapter(s). Editorial review has
deemed that any suppressed content does not materially affect the overall learning experience. Cengage Learning reserves the right to remove additional content at any time if subsequent rights restrictions require it.